Oecologia (2002) 131:227–235 DOI 10.1007/s00442-002-0885-9
PLANT ANIMAL INTERACTIONS
Jennifer S. Thaler · Richard Karban · Diane E. Ullman Karina Boege · Richard M. Bostock
Cross-talk between jasmonate and salicylate plant defense pathways: effects on several plant parasites Received: 17 July 2001 / Accepted: 3 January 2002 / Published online: 2 March 2002 © Springer-Verlag 2002
Abstract Plants are often attacked by many herbivorous insects and pathogens at the same time. Two important suites of responses to attack are mediated by plant hormones, jasmonate and salicylate, which independently provide resistance to herbivorous insects and pathogens, respectively. Several lines of evidence suggest that there is negative cross-talk between the jasmonate and salicylate response pathways. This biochemical link between general plant defense strategies means that deploying defenses against one attacker can positively or negatively affect other attackers. In this study, we tested for crosstalk in the jasmonate and salicylate signaling pathways in a wild tomato and examined the effects of cross-talk on an array of herbivores of cultivated tomato plants. In the wild cultivar, induction of defenses signaled by salicylate reduced biochemical expression of the jasmonate pathway but did not influence performance of S. exigua caterpillars. This indicates that the signal interaction is not a result of agricultural selection. In cultivated tomato, biochemical attenuation of the activity of a defense protein (polyphenol oxidase) in dual-elicited plants resulted in increased of performance of cabbage looper caterpillars, but not thrips, spider mites, hornworm caterpillars or the bacteria Pseudomonas syringae pv. tomato. In addition, we tested the effects of jasmonate-induced resistance on the ability of thrips to vector tomato spotted wilt virus. Although thrips fed less on induced plants, J.S. Thaler (✉) Department of Botany, 25 Willcocks St., University of Toronto, Toronto, M5 S 3B2 Canada e-mail:
[email protected] Fax: +1-416-9785878 R. Karban · D.E. Ullman · K. Boege Department of Entomology, University of California, Davis, CA 95616, USA R.M. Bostock Department of Plant Pathology, University of California, Davis, CA 95616, USA Present address: K. Boege, Biology Department, University of Missouri-St. Louis, 8001 Natural Bridge Road, St. Louis, MO 63121, USA
this did not affect the level of disease. Thus, the negative interaction between jasmonate and salicylate signaling had biological consequences for two lepidopteran larvae but not for several other herbivores tested or on the spread of a disease. Keywords Cross-talk · Induced defense · Jasmonate · Salicylate · Herbivory
Introduction Plants must protect themselves against many invading organisms that consume, infect, and damage their tissue in various ways (Hatcher 1995). Attack by one organism may be associated with attack by other organisms. For example, many herbivorous insects are vectors for plant disease. The physical wound created by insect feeding may itself facilitate the entry of opportunistic pathogens. Conversely, some insects such as aphids and thrips prefer the yellow color of diseased plants (Dixon 1998). Given that insect and pathogen attack are often positively associated, plants should utilize defense systems that are either effective against both types of attackers, or at a minimum that do not interfere with defenses against other attackers. Some initial work supported this idea of coordinated defenses against insects and pathogens (e.g., McIntire et al. 1980; Karban et al. 1987). However, there is a growing body of evidence demonstrating that the major biochemical pathways that mediate plant resistance to diverse attackers interact with each other and that these interactions can be negative. Induced resistance involves plant-mediated changes associated with initial attack by herbivores and pathogens that negatively influence subsequent attackers (Karban and Baldwin 1997; Agrawal et al. 1999). The jasmonate pathway (i.e., the octadecanoid pathway) and the salicylate pathway (conditioning systemic acquired resistance, SAR) are two of the biochemical response mechanisms that can be triggered by various attackers, and the components of these pathways, jasmonic and sal-
228
icylic acids, function as necessary signaling molecules that mediate such defensive responses (McConn et al. 1997; Hammerschmidt and Smith Becker 1999). Endogenous jasmonic acid (JA) induces putatively defensive phytochemicals and proteins such as proteinase inhibitors and oxidative enzymes. Some of these induced compounds have been causally linked to increased plant resistance to many insects and some pathogens (Duffey and Stout 1996; Karban and Baldwin 1997; Vijayan et al. 1998). Endogenous salicylic acid (SA) results in the production of a suite of phytochemicals that is correlated with protection from many pathogen attackers (Ryals et al. 1992; Hammerschmidt and Nicholson 1999). Jasmonate and salicylate each trigger an array of biochemical responses and products, some of which overlap, although many are distinct (Choi et al. 1994; Niki et al. 1998; Pieterse and van Loon 1999; Bostock 1999; Schenk et al. 2000). The jasmonate and salicylate pathways can negatively interact with each other at the biochemical level. Laboratory experiments have demonstrated an antagonistic interaction between pathways in tomato, tobacco, and Arabidopsis thaliana plants (Doherty 1988; Peña-Cortez et al. 1993; Doares et al. 1995a, b; Sano et al. 1995; Niki et al. 1998; Felton et al. 1999; Fidantsef et al. 1999; Gupta et al. 2000). These pathway interactions also occur in cases where plants are attacked by biotic challengers (Preston et al. 1999; Stout et al. 1999). For example, tobacco plants infected with tobacco mosaic virus displayed decreased ability to induce JA in response to mechanical wounding and increased susceptibility to the herbivore Manduca sexta (Preston et al. 1999). However, recent evidence indicates that the pathway interactions are not always negative, depending on the dose and timing of elicitation (Doherty et al. 1988; Niki et al. 1998, Thaler et al. 2002) and the response measured (Schenck et al. 2000). There is also growing evidence of salicylate and jasmonate independent plant defense responses (Pieterse and van Loon 1999) some of which can act in a compensatory manner when one pathway is disabled (Boland et al. 1998). This complexity of signaling makes the answer to the question of how cross-talk in signaling pathways affects biotic challengers less predictable. Thus, to understand the ecological significance of signal interactions in plant defense we must measure the actual level of plant resistance to herbivores rather than the level of specific defense-related compounds. Elicitors of the jasmonate and salicylate pathways are being evaluated for use in agricultural pest management and to better understand how plants coordinate defensive responses. Two elicitors were used in the current study, JA and the synthetic functional analog of salicylic acid, BTH (benzothiadiazole), to induce the corresponding response pathways. These chemical elicitors were used to identify potential interactions between the jasmonate and salicylate response pathways, and to avoid confounding influences of actual leaf damage or infection by biotic agents. Application of BTH to leaves mimics the effect of SA, apparently by interacting with the same cellular
sites of action as SA (Lawton et al. 1996; Tally et al. 1999), and does not result in a local increase in endogenous SA. Thus, all or a significant subset of the defenserelated biochemical responses induced by SA are also induced by BTH. We previously showed that simultaneous application of both JA and BTH to field grown tomato plants resulted in attenuated expression of hallmark biochemical responses to these inducers compared to plants induced with just a single elicitor (Thaler et al. 1999). Polyphenol oxidase, a JA responsive protein, had lower activity in plants elicited with both JA and BTH compared to plants elicited with only JA. Accumulation of pathogenesisrelated protein 4 mRNA (P4), a SA responsive protein, was reduced in plants elicited with both JA and BTH compared to plants elicited with only BTH. This negative interaction in the chemical expression of the two pathways corresponded to a reduction in resistance to a herbivore. We found that plants simultaneously induced with JA and BTH had compromised resistance to the caterpillar Spodoptera exigua, compared to plants induced with JA alone. The ecological and agricultural importance of plant signaling interactions will depend on the occurrence of negative interactions in other plants and the range of herbivores that are affected by the jasmonate- and salicylate-mediated responses. We evaluated this by (1) testing for signal interaction in a wild tomato variety, (2) testing the performance of four herbivore species and a pathogen feeding on cultivated tomato plants that were elicited with JA and BTH singly as well as dual-elicited plants where the two pathways showed a negative biochemical interaction and (3) testing whether induced resistance affected feeding behavior of thrips and the thrips vectored spread of tomato spotted wilt virus.
Materials and methods General methods Tomato plants were grown in 4-inch pots containing UC soil mix in a greenhouse. Plants were grown for approximately 1 month, until the four-leaf stage, after which the treatments were applied. JA and BTH were applied to plants using hand held atomizers. The JA was prepared in aqueous suspension using acetone to help disperse it in water (1% acetone in water). BTH was dissolved in water. The solutions/suspensions were sprayed to runoff onto the desired portion of the plant, shielding the rest of the leaves. Control plants were sprayed with an equal amount of water with acetone. Previous experiments demonstrated that the effects of BTH treatment are not simply due to the absence of acetone. The growth rate of S. exigua caterpillars feeding on BTH elicited plants was higher compared to caterpillars feeding on control plants that had been treated with water not containing acetone (mean±SE BTH: 0.49±0.09, Control: 0.22±0.03). Elicitation of wild tomato We tested for negative interactions between jasmonate- and salicylate- mediated induced plant resistance on a wild variety of tomato, Lycopersicon esculentum var. cerasiforme from Papantla, Veracruz, Mexico. Var. cerasiforme is the immediate ancestor of the
229 cultivated tomato and is common in Mexico, Central and South America (Rick 1995). This experiment was conducted to ensure that the signal interaction was not simply an artifact of agricultural selection. These experiments were conducted using plants in four treatments (1) control, (2) 1.5 mM JA, (3) 1.2 mM BTH and, (4) simultaneous 1.5 mM JA and 1.2 BTH (n=12 per treatment). The third youngest leaf of plants at the four-leaf stage was sprayed with the elicitors. Activity of polyphenol oxidase, performance of S. exigua, and Pseudomonas syringae pv. tomato lesion formation were assayed. Polyphenol oxidase is an enzyme induced by the jasmonate pathway in tomato plants. Polyphenol oxidase activity was chosen as a marker of the jasmonate-induced response because of its consistent systemic pattern of induction following herbivore damage and JA spray (Thaler et al. 1996). Systemic activity of polyphenol oxidase was determined 2 days following elicitor application in the terminal leaflet of the fourth leaf according to the methods described in Thaler et al. (1996). Briefly, we extracted the enzymes from weighed leaflets that were homogenized in ice-cold buffer and the homogenate was centrifuged to obtain a clarified extract for enzyme analyses. The supernatant was added to a caffeic acid solution and absorbance read at 470 nm. Polyphenol oxidase activity in the JA and JA/BTH-treated plants was compared using two-way ANOVA. Herbivore bioassay Newly hatched S. exigua caterpillars were reared on artificial diet (Southland Products) for 3–5 days before the bioassays. Two days following elicitor application, caterpillars were placed individually on a single adjacent leaflet from the fourth leaf in 90-mm petri dishes lined with moist filter paper. The caterpillars were weighed at the beginning of the experiment, and again at the end, after 2–3 days of feeding on the specified leaflets. The assays were terminated before the leaf began to deteriorate and before the caterpillars could consume the entire leaflet. The relative growth rate [(final weight–initial weight)/(initial weight × number of days)] (RGR), and gross growth efficiency [(final weight–initial weight)/leaf area consumed)] (ECI) were calculated (Waldbauer 1968). The effect of jasmonate and salicylate elicitation on each performance measure was compared using two-way factorial ANOVA. Bacterial speck disease assay Pseudomonas syringae pv. tomato (Pst) isolated from field-grown tomato plants (isolate Pst23, gift of D. Cooksey, Department of Plant Pathology, UC Riverside) was incubated at 27°C for 48–72 h on King’s B medium and colonies were suspended in sterile water. An aqueous suspension of 107 colony-forming units per milliliter was gently painted on the attached terminal leaflet of the fifth leaf of intact plants with a cotton applicator, and the plants were incubated in the greenhouse. There is a compatible interaction between this Pst strain and the plant. Seven days later the number of lesions on the inoculated leaflet was counted as a measure of disease. There is a strong correlation between the number/amount of lesions and bacterial populations (colony forming units/cm2) (C. Richael, personal communication). The effect of jasmonate and salicylate elicitation on disease was compared using two-way factorial ANOVA. Effect of signaling cross-talk on multiple herbivores In many greenhouse and field experiments, noctuid caterpillars including S. exigua and Helicoverpa zea have shown improved performance on BTH-elicited plants compared to control plants and on BTH- and JA-elicited plants compared to plants only elicited with JA (Stout et al. 1999; Thaler et al. 1999). We examined if a range of insects are affected by the jasmonate and salicylate signal
interaction, by testing the effects of JA and BTH application on spider mites, thrips, hornworms, cabbage loopers, and bacterial speck disease. The cultivated tomato (L. esculentum cv. New Yorker) was used for these experiments. A JA and BTH elicitor regime where the chemical signaling interactions are most consistent was chosen (Thaler et al. 1999, unpublished data). These experiments were conducted using plants in four treatments (1) control, (2) 1.5 mM JA, (3) 1.2 mM BTH and, (4) simultaneous 1.5 mM JA and 1.2 mM BTH. The third leaf of plants at the fourleaf stage was sprayed with the elicitors. Nine sets of 48 plants were used in total for these experiments. Twelve replicates of each treatment were conducted during each experiment. Polyphenol oxidase activity was assayed on four of the nine sets of plants: one set that was used for a thrips, spider mite and Pst assay, one set that was used for a hornworm assay, one set that was used for a hornworm and a thrips assay, and one set for the cabbage looper assay. Polyphenol oxidase activity was measured to ensure that the negative biochemical signal interaction had occurred between the jasmonate and salicylate pathways as in previous experiments. In the four sets of plants where polyphenol oxidase was measured, the adjacent leaflet of the fourth leaf was excised from all 12 plants in each of the four treatments. There was statistically significant lower polyphenol oxidase activity in the plants that were elicited with JA and BTH compared to plants that were elicited with JA alone. This indicates that the chemical crosstalk had occurred (see Results). The effects of this signal interaction on plant resistance to western flower thrips (Frankliniella occidentalis), two-spotted spider mites (Tetranychus urticae), tobacco hornworm caterpillars (Manduca sexta), cabbage looper caterpillars (Trichoplusia ni) and the causal agent of bacterial speck disease (Pst) were tested. These are organisms that naturally feed on or infect tomato plants in the field. Thrips, spider mites, cabbage loopers and Pst are generalists; hornworm caterpillars are specialists on Solanaceous plants. Thrips and spider mites are cell content feeders, hornworm and cabbage looper caterpillars are leaf chewers, and Pst is a bacterium causing lesions on leaves, stems and fruits. These organisms were obtained from laboratory colonies. The terminal or adjacent leaflet of the fourth leaf was collected 3 days after JA and BTH applications for the herbivore bioassays. For the thrips, hornworm, and cabbage looper caterpillar assays, leaflets were placed in a petri dish lined with moist filter paper and sealed with Parafilm. Either individual second instar thrips larvae or adult thrips were placed on each leaflet and allowed to feed for 3 days after which survivorship and the amount of leaf area eaten (mm2) was quantified using an acetate grid. Single newly hatched hornworm or cabbage looper caterpillars were placed on each leaflet and allowed to feed for 3 days after which survivorship and RGR were measured. Because the spider mite bioassays lasted longer, the cut end of the petiole was covered with a moist cotton ball to maintain leaf water content. Three female spider mites were placed on each leaflet and allowed to reproduce for 10 days, after which the number of eggs, immatures, and adults was counted using a stereo microscope. Pst was inoculated onto the attached terminal leaflet of the fourth leaf and lesions counted 7 days later using the same methods as in the wild tomato experiment described above. Using the nine sets of plants, one trial of the cabbage looper and Pst assay, two trials of the spider mite and hornworm assay, and six trials of the thrips assay were conducted. Three trials were conducted using immature thrips and three trials were conducted using adult thrips. Because the immature and adult thrips responded in the same way to the elicitor treatments, the results were pooled. Bioassays were conducted on excised leaflets and therefore some sets of plants were used for challenges against two or three herbivore species (hence 12 assays were conducted with 9 sets of plants).
Vector transmitted disease We examined the effect of induction of the jasmonate pathway on the development of disease caused by tomato spotted wilt virus
230 (TSWV), a virus frequently vectored by thrips. We tested whether expression of the jasmonate-regulated responses or the reduced feeding of thrips on jasmonate-induced plants would affect viral transmission and replication. Ten-day-old greenhouse-grown tomato plants (cv. Celebrity) were divided into two groups of 25 plants each, half that were treated with 0.5 mM JA, the other half left as controls. Two days later, polyphenol oxidase activity was measured on an excised leaflet from eight control and eight sprayed plants to ensure that the jasmonate pathway was induced. The plants were then transferred to the laboratory and randomly interspersed in a Plexiglas cage (2×0.9×0.9 m). Approximately 900–1,000 viruliferous thrips, obtained from a colony feeding on Emilia sonchifolia plants infected with TSWV, were released into the cage containing 25 induced and 25 control plants. Because all of the plants were in a single cage, the feeding choices of the thrips could have been influenced by neighboring plants. Four days after the thrips were released, new leaflets were collected to measure polyphenol oxidase activity in eight control and eight jasmonate-induced plants. A different subset of plants was assayed from the ones assayed for polyphenol oxidase activity prior to thrips release. The amount (mm2) of feeding damage on all plants was quantified using an acetate grid. The plants were then fumigated to remove all thrips and allowed to grow for 3 weeks to allow the virus to develop. Enzyme-linked immunosorbent assay was conducted to quantify the presence of TSWV using methods similar to Ullman et al. (1993). TSWV infected E. sonchifolia and tomato were used as positive controls, while healthy plants of the same age were used as negative controls. The samples were read at 405 nm 60 min after the start of the reaction. Plant samples with optical density readings greater than two standard deviations beyond the mean of the negative controls were considered positive. A second trial of this experiment was conducted using similar methods except that 15 control and 15 induced plants were used and plates were read 30 and 60 min after the start of the reaction. Polyphenol oxidase activity and amount of thrips feeding damage were analyzed using two-way ANOVA with induction treatment and trial as the main effects. Virus titers were analyzed using MANOVA with time 30 and 60 as the dependent variables.
Results and discussion Elicitation of wild tomato JA stimulated and BTH reduced the activity of polyphenol oxidase compared to controls. Polyphenol oxidase activity was intermediate in the dual- elicited plants (Fig. 1 a, Table 1). This result confirmed the negative effect of SAR activation on the jasmonate pathway that we observed previously in the cultivated tomato. Neither JA or BTH affected the RGR or ECI of S. exigua (Fig. 1b, c, Table 1). Surprisingly, the number of Pst lesions was increased by BTH and not affected by JA (Fig. 1d, Table 1). We observed much higher levels of disease on these plants compared to what is typically observed on cultivated tomato plants in similar greenhouse trials (see data below). Some features of the negative chemical signaling interaction between JA and SA were seen in the wild tomato. Polyphenol oxidase activity was reduced on dualtreated plants compared to JA-treated plants alone. However, this did not increase performance of S. exigua on dual- treated plants, indicating discordance between biochemical responses and biological resistance. The Pst results were surprising, with elicitation of SAR by BTH dramatically increasing lesion number. This is contradic-
Fig. 1a–d The effect of JA (1.5 mM) and BTH (1.2 mM) treatment on the wild tomato, Lycopersicon esculentum cv. cerasiforme. a Activity of polyphenol oxidase (∆OD/g/min) (PPO), performance of the herbivore Spodoptera exigua measured as b relative growth rate [(final weight–initial weight)/(initial weight×number of days)] (RGR), and c gross growth efficiency [(final weight-initial weight)/mm2 leaf area consumed] (ECI), and d disease caused by the pathogen Pseudomonas syringae pv. tomato (mm2 of lesions formed) was measured. Bars indicate mean±SE
tory to our results with the cultivated tomato and may represent differential responses to BTH in the wild tomato. Thus, although further studies are required, negative signal interactions between jasmonate and salicylate signaling seem likely in wild systems (Preston et al. 1999).
231 Table 1 Elicitation of wild tomato. ANOVA analysis of polyphenol oxidase activity, the performance of herbivores and the levels of disease on plants in four treatments: Control, JA, BTH and BTH/JA. The elicitor (1.5 mM JA and 1.2 mM BTH) was simultaneously applied to plants. Polyphenol oxidase activity (PPO) was calculated as the ∆OD/g/min. The relative growth rate (RGR) of Spodoptera exigua caterpillars is the (final weight–initial weight)/(initial weight) and gross growth efficiency (ECI) is (final weight–initial weight)/ (leaf area consumed). Disease caused by Pseudomonas syringae pv. tomato was measured as the mm2 of lesions Experiment
Factor
df
F
P
PPO activity
JA BTH JA×BTH Error JA BTH JA×BTH Error JA BTH JA×BTH Error JA BTH JA×BTH Error
1 1 1 44 1 1 1 42 1 1 1 42 1 1 1 44
10.885 7.209 2.331 – 1.497 0.023 0.037 – 1.944 1.887 0.025 – 2.05 6.186 2.017 –
0.002 0.010 0.134 – 0.228 0.879 0.849 – 0.171 0.177 0.852 – 0.159 0.017 0.163 –
RGR of S. exigua
ECI of S. exigua
Pst lesions
Table 2 Generality of antagonism. ANOVA analysis of the performance of herbivores on plants in four treatments: Control, JA, BTH and BTH/JA. The elicitor (1.5 mM JA and 1.2 mM BTH) was simultaneously applied to plants. The RGR of hornworm and cabbage looper caterpillars is the (final weight–initial weight)/(initial weight). Thrip damage is measured as mm2 of damage. The number of spider mites is the sum of adults, immatures, and eggs. Disease caused by P. syringae pv. tomato was measured as the mm2 of lesions. All interactions were included in the model; nonsignificant interactions are not reported Experiment
Factor
Hornworm RGR
JA 1 BTH 1 Trial 1 JA×BTH 1 JA×BTH×Trial 1 Error 85 JA 1 BTH 1 JA×BTH 1 Error 42 JA 1 BTH 1 Trial 5 JA×BTH 1 Error 261 JA 1 BTH 1 Trial 1 JA×BH 1 Error 87 JA 1 BTH 1 JA×BTH 1 Error 44
Cabbage looper RGR
Thrips damage
Number of spider mites
df
F
P
1.555 1.827 9.095 4.041 4.423 – 9.956 43.457 0.290 – 9.378 0.095 5.042 0.107 – 4.680 0.907 0.002 0.132 – 8.264 4.216 1.518 –
0.216 0.180 0.003 0.048 0.038 – 0.003
